Income Diversification and Poverty Reduction
in the Northern Uplands of Vietnam
Nicholas Minot
Research Fellow
International Food Policy Research Institute
2033 K St. N.W.
Washington, D.C. 20036
Telephone: 202 862-8199
Email: [email protected]
Paper prepared for presentation at the American Agricultural Economics Association Annual Meeting,
Montreal, Canada, July 27-30, 2003
Copywrite 2003 by Nicholas Minot. All rights reserved. Readers may make verbatim copies of this document
for non-commerical purposes by any means, provided that this copyright notice appears on all such copies.
Abstract:
In the context of the economic development, income diversification is sometimes defined as the
process by which households switch from growing low-value staple food crops to growing a mix of
food crops and higher-value commercial crops (crop diversification) and from farming to non-farm
activities (non-farm diversification). The literature on income diversification has measured trends,
identified determinants, and speculated on the constraints to diversification, but there has been
relatively little analysis of the contribution of diversification to income growth. This paper uses
household survey data from 1993 and 1998 to quantify the contribution of crop diversification and
non-farm diversification to the growth of household income in the northern upland region of Vietnam.
We find that rural incomes have increased substantially over this period, but non-farm income has
increased at the same rate as farm income. Poor households are particularly dependent on crop
income growth, while higher-income households rely more on non-farm diversification to increase
their incomes. Crop diversification has made a non-negligible contribution to rising living standards
in the northern uplands, but the contribution of yield growth is substantially greater. Crop income
growth among poor households is based heavily on yield increases, while income growth among
richer households is associated with area expansion and crop diversification. These results have
implications for policies and investment to improve rural living standards.
Key words: Income growth, diversification, Vietnam
Acknowledgements: Funding for this research was provided by the Japanese Bank for International
Cooperation. The author would like to thank Michael Epprecht for his assistance in coordinating the
field work associated with this project and Reno Dewina for her assistance in processing the data and
formatting the tables.
1
Introduction
Vietnam has maintained an annual economic growth rate of 6-8 percent since the early 1990.
Furthermore, this economic expansion has had concrete benefits for low-income households, reducing
the rural poverty from 66 percent in 1993 to 45 percent in 1998 (Joint Government-World Bank-NGO
Working Group, 2000). Part of the reduction in poverty is due to higher yields of rice and other
crops, which have allowed Vietnam to become the second largest rice exporter with no expansion in
rice area and no reduction in domestic consumption (Minot and Goletti, 2000). In addition,
Vietnamese farmers have diversified into higher-value crops. Vietnam has become the second largest
coffee producer in the world, and production and export of fruits and vegetables have risen
dramatically over this period. And part of the income growth and poverty reduction is undoubtedly
due to diversification into non-crop activities such as aquaculture, livestock raising, and non-farm
activities.
The importance of each of these factors in rural income growth and poverty reduction has
implications for policy and public investment. If most rural income growth comes from technological
change which increases yields, then investments in agricultural research and extension deserve
priority1. If income growth derives largely from crop diversification, then attention needs to be
focused on agricultural credit, transportation, and market information to facilitate this process. And if
income growth or poverty rises mostly due to diversification into non-farm activities, then perhaps the
focus should be on training, electrification, and commercial credit to stimulate non-farm employment
growth.
It should be noted, however, that there may be a trade-off between raising rural incomes in
general and poverty alleviation. For example, the factors that contribute to income growth for poor
households may differ from those that drive income growth among other households. Thus, it is
important to look at the sources of growth for different income groups.
1
In theory, investment decisions should be made based on the size of the marginal contribution to
growth from alternative investments. However, since this information is generally not available, a reasonable
approximation is to allocate investment among growth strategies according to the average contribution to
economic growth.
Income diversification and poverty reduction in Vietnam
Page 1
This study focuses on the patterns of income growth in the northern uplands, consisting of 14
provinces. This region is characterized by rugged topography, poor infrastructure, a relatively low
population density, and a high proportion of ethnic minority households. Thus, it is not surprising that
the region has the highest poverty rate in the country. According to one study, the ten poorest
provinces in Vietnam are located in the northern uplands (Minot and Baulch, 2002).
In light of the lack of information on the sources of growth, this study will focus on the
following questions:
•
What is the relative importance of various sources of rural income growth in Vietnam:
yield increases, crop price increases, diversification into high-value crops, growth in
livestock and aquaculture, or the shift toward non-farm activities.
•
What is the relative importance of these factors in the dramatic reduction in rural poverty
in Vietnam?
•
How do the sources of income growth and poverty reduction vary by region and by type
of household?
Section 2 provides a brief review of studies of the causes and effects of income
diversification. Section 3 describes the data and methods used to decompose income growth and
measure the contribution of diversification. In Section 4, we describe the patterns of rural growth in
income and expenditure between the two surveys. Section 5 decomposes rural income growth for
different groups of households, focusing on the contribution of diversification to growth. And Section
6 summarizes the results and discusses some policy implications.
2
Background
Crop diversification is a somewhat ambiguous concept, being associated with both
subsistence farmers and commercialization. In some contexts, poor, semi-subsistence farmers grow a
wide variety of food crops in order to reduce the risk of weather- or disease-related crop failure or to
meet diverse home consumption needs. This is particularly relevant for farmers in remote regions,
growing crops under rain-fed conditions where rainfall is not reliable, as in many parts of sub-Saharan
Africa. In contrast, in areas where poor farmers are heavily dependent on a staple crop (such as the
rice economies of Asia), crop diversification is associated with the process of commercialization, in
which farmers begin to combining food crops with higher-value commercial crops (Delgado and
Income diversification and poverty reduction in Vietnam
Page 2
Siamwalla, 1997). Joshi et al (2002) estimate crop diversity using state-level time-series data from
India and find that it is negatively related to rainfall but positively related to road density, per capita
income, and urbanization. In Cote d’Ivoire, the 1994 currency devaluation increased the incentive to
diversify into cocoa, cotton, and other export crops, but richer households were better able to take
advantage of this opportunity, presumably due to greater liquidity (Barrett et al, 2001). A study of
Vietnam found that diversification away from rice is associated with small farms, small irrigated area,
and more education (Pederson and Annou, 2002).
Income diversification is usually defined in terms of diversification into non-agricultural (or
sometime non-crop) income sources. Research on this topic has examined the importance of rural
non-farm income, its relationship with income and other household characteristics, the efficiency of
small enterprises and their prospects for survival, and the role of farm-non-farm linkages (see for
example Lanjouw and Lanjouw, 2001; Delgado et al, 2000; Reardon, 1997; and Hazell and Roell,
1983). Reardon (1997) summarizes the results of 27 studies of rural non-farm employment in subSaharan Africa, finding that a) wage labor is more important than enterprise income, b) non-farm rural
income is more important near cities in areas with high population density and good infrastructure,
and c) non-farm rural income is more common among higher-income rural households. A study of
three African countries finds that income diversity is greater among high-income rural households,
contradicting the idea that diversification is a risk-reduction strategy. Unskilled labor income is more
common among poor households, but other types of non-farm income are more prevalent among
richer households (Barrett et al, 2000). In Ecuador, the share of rural income from non-farm
enterprises is positively correlated with education, electrification, proximity to markets, and the value
of crop production per hectare (Escobal, 2001). And a study of the northern uplands in Vietnam
asked farmers about the main constraints to income diversification and poverty reduction. The
responses included lack of capital, shortage of paddy land, poor access to markets, poor irrigation
infrastructure, and insufficient education (Henin, 2002).
Income diversification and poverty reduction in Vietnam
Page 3
3
Data and methods
What is income diversification and how do we measure it? Before analyzing the patterns and
trends in income diversification in the north mountain region of Vietnam, it is important to address
these two questions. In this section, we briefly describe the data used in this section, the definitions of
diversification adopted, and the measures of diversification used.
3.1
Data
This report uses two national household surveys to examine the patterns and trends in income
diversification. The first Vietnam Living Standards Survey (VLSS) was carried out in 1992-93 by the
State Planning Committee and the General Statistics Office, with financial contributions from the
United Nations Development Programme (UNDP) and the Swedish International Development
Agency (SIDA) and technical assistance from the World Bank. The 110-page questionnaire was
administered to 4800 households between October 1992 and October 1993 (see SPC/GSO, 1994 for
more details). Because the bulk of the data collection took place in 1993, we will refer to it as the
1993 VLSS.
The second Vietnam Living Standards Survey was conducted in 1997-98 by the General
Statistics Office, with financial support from the UNDP and SIDA and technical assistance from the
World Bank. The household questionnaire covers the same topics as the 1993 VLSS, with only slight
changes in the questions and format. The sample contains 6000 households, including most of the
households from the 1993 VLSS, as well as additional households to provide better coverage of urban
areas and the Central Highlands2. The data collection took place between December 1997 and
December 1998, so we refer to this survey as the 1998 VLSS.
3.2
Definitions of diversification
The term “income diversification” has been used to describe many different concepts. One
definition of income diversification relates to the number of income sources and the balance among
2
Because the 1993 VLSS sample was designed to be proportional to the population, the sample for
the sparsely populated Central Highlands was just 128 households. Two clusters from the 1993 VLSS sample in
the Red River Delta were dropped and 1290 households were added.
Income diversification and poverty reduction in Vietnam
Page 4
them (see Joshi et al, forthcoming). Diversification is sometimes defined as the process of switching
from subsistence production of staple crops to commercial production of a wider range of agricultural
commodities and to non-farm activities. A number of studies focus on diversification from farm to
non-farm activities. Finally, diversification is often used to refer to the process by which farmers
switch from low-value crops and activities to higher-value crops and activities.
The measures of diversification discussed above are static measures in that they quantify the
degree of income diversification (however defined) at one point in time. We are also interested in
measuring the process of diversification over time. In particular, we would like to measure the
contribution of diversification to income growth. If we simply calculate the average value of crop
production per hectare at two points in time, we are including the effect of price changes and yield
changes to income growth. Thus, in order to assess the contribution of diversification, it is necessary
to develop a way of decomposing income growth into various components, one of which is
diversification.
3.3
Decomposing rural income growth
We can measure the contribution of income diversification to income growth by decomposing
growth into increases in crop income and increases in other income, then separating crop income
growth into four components: changes in yield, changes in real prices, changes in total area sown, and
crop diversification, where crop diversification is the effect of reallocating land among crops on
income, holding prices, yields, and total area constant.
We start with an expression for total net revenue in terms of crop income and non-crop
income. Crop income can be rewritten as the product of the area planted, the average yield, and the
average value per kilogram. Area, in turn, can be divided up into total area and the shares allocated to
each crop:


R = ∑ A i Yi Pi + NCY =  ∑ a i Yi Pi ∑ A i + NCY
i
 i
 i
where R = net revenue from crop production
Yi = yield of crop i (production per unit of area sown)
Pi = real net income from crop i per unit of output
Ai = area sown with crop i
Income diversification and poverty reduction in Vietnam
Page 5
a = share of crop sown area allocated to crop i
NCY = non-crop income
Next, we take the total derivative of both sides:

 
 
 

dR ≅  ∑ ai Yi Pi d  ∑ Ai  +  ∑ Ai d  ∑ ai Yi Pi  + dNCY
 i
  i
  i
  i

The second term on the right-hand side can switched from the change in a sum to the sum of changes:

 

dR ≅  ∑ ai Yi Pi d  ∑ Ai  + ∑ Ai ∑ d (ai Yi Pi ) + dNCY
i
 i
  i
 i

 

dR ≅  ∑ ai Yi Pi d  ∑ Ai  + ∑ Ai ∑ ai Yi dPi + ∑ Ai ∑ ai Pi dYi + ∑ Ai ∑ Yi Pi dai
i
i
i
i
i
 i
  i
 i
+ dNCY
The five terms on the right-hand side of the equation can be described as follows:
•
The first term on the right side represents the change in crop revenue due to the change in
total area allocated to crops. The expression ΣaiYiPi is the weighted average net income
per hectare, where the weights are the proportion of total sown area allocated to each crop
(ai).
•
The second term on the right side is the change in crop revenue attributable to the change
in real prices of the crops. The first summation is the total area, while the second
represents the change in average gross revenue per hectare due to price changes.
•
The third term is the change in gross crop revenue due to changes in yields. The first
summation is the total area, while the second is the change in average gross revenue per
hectare due to yield increases.
•
The fourth term on the right side represents the change in agricultural revenue due to crop
diversification, that is, the shift in the percentage allocation of land among crops. Again,
the first summation is the total area, while the second is the change in average gross
revenue due to shifts in the allocation of land among the crops. This term will be zero if
there is no reallocation of land among crops (dai = 0 for all crops). It will also be zero if
the revenue per hectare is the same for all crops, since Σdai = 0.
•
And the fifth term is simply the change in non-crop income. Non-crop activities can be
further disaggregated into livestock activities, fishery activities (including aquaculture),
forestry, wages, non-farm enterprise activities, transfers, and other income. These
different terms have been combined to simplify the exposition.
Thus, the contribution of crop diversification to overall growth in crop income is measured by
calculating the income change that would occur if cropland were reallocated among crops the way it
Income diversification and poverty reduction in Vietnam
Page 6
actually was between 1993 and 1998, but yield, prices, and total area remained constant. Dividing
both sides of the equation by the overall change in income (dR) will give the proportional contribution
of each component to overall growth. Naturally, this decomposition can be carried out for any region,
income group, or any other category of households.
In interpreting the results, there are three qualifications that should be kept in mind. First, the
decomposition is only approximate because there is an interaction term that reflects, for example, the
effect of higher yields on the additional area planted.
Second, as mentioned in Section 2.2, the VLSS questionnaire does not link some inputs (hired
labor, equipment rental, and storage costs) to a specific crop, so these costs are allocated in proportion
to the value of output. Furthermore, the VLSS does not provide information on the cost of planting
tree crops unless they were planted in the year of the survey.
Third, for fruit trees and industrial tree crops, the VLSS gives respondents the choice of
expressing the total area and the productive area in hectares or in the number of trees. Since the
decomposition of crop income relies on area estimates in hectares, we need to impute the area of tree
crops for households that only gave the number of trees. This was done by estimating the average
yield (output per hectare of productive land) based on the responses of those who gave area figures in
hectares. For farms that only gave the number of trees, the productive area in hectares was calculated
by dividing household output by the average yield for that crop and that region (or using a national
average yield if necessary). With information on the productive area and the productive number of
trees, the unproductive area was imputed from the number of unproductive trees, assuming that the
tree density was the same for productive and unproductive areas.
4
Sources of income and patterns in growth
How do rural households in the Northern Upland earn their living? Based on the results of
the 1998 VLSS, virtually all rural households in the Northern Uplands (100 percent of the sample)
have some crop production, and it accounts for 38 percent of net household income. The most
widespread crops are rice (grown by 91 percent of the rural households), maize (62 percent), water
morning glory (58 percent), cassava (49 percent), and bananas (47 percent). In value terms, rice is
Income diversification and poverty reduction in Vietnam
Page 7
the most important crop, accounting for 46 percent of the total net value of crop production, followed
by maize (15 percent). Tea, cassava, and litchi/longan/rambuttan, each account for 5-10 percent of
the value of crop production
How have the materials standards of living of rural Vietnamese households changed between
1993 when the first VLSS was carried out and 1998 when the second one was. Two measures of
standard of living are used: per capita consumption expenditure and per capita net income. Per capita
expenditure is defined as the sum of purchases on consumption goods, the value of food produced by
the household for home consumption, and the rental value of consumer durables and housing. These
variables were calculated by the team of analysts at the GSO and the World Bank that first processed
the results of each survey. The expenditure variables are constructed to be comparable with each
other3. Per capita net income was calculated for this report. Both are expressed in constant
Vietnamese dong at prices of January 2002.
The results confirm that the rapid economic growth recorded in national accounts has
translated into concrete gains for rural households. Between 1993 and 1998, estimated per capita
expenditure in rural areas rose 32 percent, while estimated per capita income in rural areas rose 80
percent (see Table 1 and Table 2). It is not clear whether the higher estimated income growth reflects
reality or not.4.
Female-headed households do not seem to be disadvantaged, either in terms of the level of
standard of living or in terms of participation in the gains from economic growth. If anything, femaleheaded households may be somewhat better off, at least in terms of material well-being. On the other
hand, both income and expenditure data indicate that ethnic minorities are considerably percent poorer
than the majority Kinh/Hoa. The expenditure data (which is probably more reliable) suggest that the
3
Some minor types of expenditure were excluded from the 1998 VLSS analysis because data on those
expenditures had not been collected as part of the 1993 VLSS.
4
One possibility is that incomes have increased more rapidly than expenditure, implying a large
increase in household savings and/or investment. A second possibility is that these results are accurate for the
years of the survey, but reflect annual volatility in income rather than a trend. A third possibility is sampling or
non-sampling error contribute to this discrepancy. It should be noted that income and expenditure were not
constructed to match each other. For example, expenditure includes the rental equivalent of housing and
consumer durables because it was designed to measure the standard of living, but income excludes this
component because the purpose was to focus on sources of income. Thus, it is not possible to interpret
differences between income and expenditure as savings.
Income diversification and poverty reduction in Vietnam
Page 8
standard of living of ethnic minorities is 30-38 percent below that of Kinh/Hoa households and that it
has risen but more slowly than that Kinh/Hoa households.
If we divide the sample for each year into expenditure quintiles5, then the growth in per capita
expenditure is higher for the richer quintiles. The patterns for income are less clear, but growth in per
capita income over 1993-98 for the poorest quintile (40 percent) is substantially than growth in the
other four quintiles (72-101 percent). These results indicate that the gap between rich and poor is
widening, even though the standard of living of the poor is rising.
Table 1.
Summary changes in expenditure for rural households
Per capita expenditure
1993
1998
--(1000 VND/year)--
Sex of household
Male
1,837
Female
2,053
Ethnicity
Kinh/Hoa
1,972
Minority
1,363
Expenditure Category
Poorest
930
2
1,341
3
1,724
4
2,271
Richest
3,866
Region
North Upland
1,529
Red River Delta
1,759
North Central Coast
1,616
South Central Coast
1,988
Central Highland
1,751
Southeast
2,475
Mekong River Delta
2,250
All rural areas
1,886
Source: Analysis of the 1993 and 1998 VLSS.
Change
1993-1998
(percent)
2,429
2,703
32
32
2,637
1,704
34
25
1,206
1,774
2,297
3,155
5,432
30
32
33
39
41
2,008
2,560
2,342
2,462
2,231
3,909
2,530
2,488
31
46
45
24
27
58
12
32
5
The expenditure categories are defined in terms of national quintiles for that year. For example, the
first quintile in the 1993 data includes rural households that were below the 20th percentile nationally in 1993.
Because the quintiles are defined nationally, each quintile will not necessarily represent 20 percent of the rural
population. Because the quintiles are defined for each year, expenditure ranges for each quintile vary across
years.
Income diversification and poverty reduction in Vietnam
Page 9
Table 2. Summary changes in income for rural households
Per capita income_
1993
1998
--(1000 VND/year)-Sex of household
Male
1,565
Female
1,622
Ethnicity
Kinh/Hoa
1,641
Minority
1,197
Expenditure Category
Poorest
970
2
1,122
3
1,624
4
1,832
Richest
3,178
Region
North Upland
1,474
Red River Delta
1,593
North Central Coast 1,172
South Central Coast 1,142
Central Highland
1,318
Southeast
2,163
Mekong River Delta
1,946
All rural areas
1,578
Source: Analysis of the 1993 and 1998 VLSS.
Change
1993-1998
(percent)
2,818
2,913
80
80
3,029
1,833
85
53
1,359
2,159
2,787
3,691
6,017
40
92
72
101
89
2,252
2,618
2,375
2,480
3,104
4,081
3,186
2,838
53
64
103
117
136
89
64
80
The regional results also show some divergence. According to expenditure data, the
households in the Southeast have gained the most over the 1993-98 period (58 percent growth), while
those in the Mekong Delta have gained the least (12 percent). Looking at the income data, households
in the Central Highlands had the fastest growth over 1993-98, probably reflecting the expansion in
coffee production over this period. The Northern Uplands experienced average growth in per capita
expenditure over both periods. It should be kept in mind that the in-migration of poor households into
a region will lower the average growth rate.
5
Decomposition of rural income growth
This section measures the contribution of different components of income to overall growth in
rural income. First, we examine the contribution of diversification from crop production into highervalue activities such as livestock, fisheries, non-farm enterprises, and wage income. The contribution
of a given activity is calculated as the change in income from that source as a percentage of the
overall change in income. Second, we explore the contribution of crop diversification to the growth
in overall net revenue from crop production. In this case, diversification is measured as the increase
Income diversification and poverty reduction in Vietnam
Page 10
in income that can be attributed to the reallocation of land among crops, holding constant yields,
prices, and total area cropped. The method is described in more detail in Section 2.4.
5.1
Income diversification
Income diversification in the Northern Uplands
Among rural households in the Northern Uplands, net income increased from 6.9 million
VND per household per year in 1993 to 11.0 million VND in 1998 (expressed in constant terms at
January 2002 prices). These imply growth rates of 59 percent over 1993-19986. The composition of
income has changed slowly over time, with agriculture and enterprise income becoming less
important and wage and forestry income becoming more important. Income from livestock, fisheries,
and transfers remained roughly constant as a proportion of the total. Although the importance of crop
income has declined, crop and livestock income still represent over half of the rural income in the
Northern Uplands (see Table 3).
Table 3. Contribution to overall growth of each source of income
in the rural Northern Uplands
Source
Net income
1993
1998
(1000 VND/hh/year)
Crops
3,249
5,065
Livestock
785
1,097
Fisheries
214
310
Forestry
137
380
Enterprise
1,309
1,941
Wages
539
982
Transfers
680
1,146
Other
14
64
Total
6,928
10,985
Source: Analysis of the 1993 and 1998 VLSS.
Share of income
1993
1998
--(percent)-47
46
11
10
3
3
2
3
19
18
8
9
10
10
0
1
100
100
Table 4 shows the growth rates for income from each source and the contribution of each type
of income to overall income growth. For example, between 1993 and 1998 crop income rose from 3.2
million VND to 5.1 million VND in 1998, while total income rose from 6.9 million VND to 11.0
million VND. Thus, the increase in net income from crop production (1.9 million VND) contributed
45 percent of the increase in total net revenue (4.0 million VND) over this period.
6
These growth rates differ somewhat from the per capita income growth rates reported in Table 2
because of changes in the average household size.
Income diversification and poverty reduction in Vietnam
Page 11
Table 4. Growth of income from each source and contribution
to overall growth of each source in the rural Northern Uplands
Source
Contribution
Growth
to total growth
1993-98
1993-98
----(percent)----Crops
56
45
Livestock
40
8
Fisheries
45
2
Forestry
178
6
Enterprise
48
16
Wages
82
11
Transfers
69
11
Other
344
1
Total
59
100
Source: Analysis of the 1993 and 1998 VLSS.
Applying similar calculations to other activities, it appears that the growth in enterprise
income accounts for 16 percent of the overall growth. This is somewhat surprising in light of the
results presented in Section 5.1 showing that the proportion of households with enterprise income fell
substantially between the 1993 and 1998 VLSS surveys. Given that the proportion of rural
households in the region with enterprise income has fallen but the total enterprise income increased 48
percent in real terms, the data suggest that the small enterprise sector is undergoing some form of
consolidation, as hypothesized in Section 5.1. In other words, fewer household operate enterprises
but the average size of the enterprises is rising.
The growth in wage income and in transfers each account for 11 percent of the overall growth
in the net revenue of rural households in the Northern Uplands over 1993-98, while growth in
livestock income accounts for 8 percent of the total growth. Although forestry income shows the
fastest growth among the eight income sources, its contribution to overall growth is still relatively
small (6 percent) because it is such a small source of income (see Table 4).
If we define livestock, fisheries, and forestry as high value agricultural activities, then the
growth in high value agricultural activities accounts for 16 percent of the growth in overall income. If
we consider non-farm enterprises and wage labor together, then growth in non-farm activities
represents 27 percent of the overall growth in income.
Income diversification and poverty reduction in Vietnam
Page 12
Income diversification in other regions
How does the contribution of each income source to overall growth in rural income vary
across regions? Over the period 1993-98, the contribution of crop production to income growth
varied from 30 percent in the North Central Coast and the Southeast to 75 percent in the Central
Highlands. The small contribution in the Southeast is due to the high level of urbanization and the
availability of non-farm employment which mean that wages are an important source of income
growth in this region. The large contribution of crop income in the Central Highlands is linked to the
boom in coffee production during the mid-1990s. In the other three regions (the two Deltas and the
South Central Coast), crop production accounts for 47-58 percent of the overall growth (see Table 5).
In spite of the variation in the contribution of crop production growth to overall growth, it is
noteworthy that crop production is the most important source of rural income growth in all seven
regions of Vietnam. The second largest contributor to rural income growth varies across regions. In
the Northern Uplands, the Red River Delta, the Central Highlands, and the Mekong River Delta, nonfarm enterprise income is the second largest contributor to rural income growth. In the South Central
Coast and the Southeast, wages are the second largest contributor (see Table 5). In contrast, for urban
households (not shown in the table), the sources of income growth are split almost exactly between
wages, 50 percent, and non-farm enterprise income, 49 percent.
Table 5. Contribution to overall growth of each income source in rural areas by region
period 1993-1998
Source
N.
Uplands
Red
N.C.
S.C.
C.
South- Mekong
River
Coast
Coast
Higheast
River
Delta
lands
Delta
---------(percent of overall income growth)--------Crops
45
47
30
55
75
30
58
Livestock
8
-1
7
8
10
10
9
Fisheries
2
4
3
0
1
1
8
Forestry
6
0
5
2
0
2
-2
Enterprise
16
21
7
2
11
10
12
Wages
11
16
17
30
1
26
7
Transfers
11
12
28
3
1
18
11
Other
1
1
3
-1
1
3
-3
Total
100
100
100
100
100
100
100
Source: Analysis of the 1993 and 1998 VLSS.
Rural
average
48
7
3
2
11
15
13
1
100
Income diversification by income group
This section compares rice and poor households in terms of the income growth over 1993-98
and the contribution of each income source to overall growth over this period. There are two ways to
Income diversification and poverty reduction in Vietnam
Page 13
define the income categories. The top half of Table 6 divides households according to the national
quintile of per capita expenditure for that year. In other words, it compares the poorest 20 percent in
1993 with the poorest 20 percent in 1998. The average income in the poorest quintile rose 46 percent
while the average income in the richest quintile rose 125 percent. This implies that the gap between
the poor and rich has widened, even though the poor have benefited from solid income growth.
The bottom half of Table 6 makes use of the fact that the 1998 VLSS contains many of the
same households that were in the 1993 VLSS. In this section, we limit our analysis to households that
were in both surveys, of which there are 4,299 in total and 635 in the rural Northern Uplands. By
doing this, we can classify households according to their standard of living in the base year 1993,
regardless of how much their incomes changed over the period 1993-98. More specifically, the
sample is divided into quintiles according to the value of per capita expenditure in 1993. When the
households are organized in this way, the growth in net income over 1993-98 does not show any
consistent pattern across expenditure quintiles. The growth in net income among the poorest quintile
was 57 percent, compared to 46 percent in the highest quintile. The paradoxical result is that the gap
between the poor and the rich in the rural areas of Vietnam has widened, but those who were poor in
1993 gained as much as those who were rich in 1993, on average.
Table 6.
Income growth between 1993 and 1998 by expenditure category
Expenditure
category in
each year
Poorest
2
3
4
Richest
Expenditure
category in
1993
Net Income
1993
1998
---(1000 VND/hh/year)--4,715
6,8837
6,238
11,219
8,691
12,128
8,765
16,606
11,385
25,624
Growth
1993-98
(percent)
46
80
40
89
125
Net Income
Growth
1993
1998
1993-98
---(1000 VND/hh/year)--(percent)
Poorest
4,823
7,5577
57
2
6,296
9,520
51
3
8,808
13,836
57
4
9,113
14,818
63
Richest
11,385
16,656
46
Source: Analysis of the 1993 and 1998 VLSS. Top half of the table includes all households. Bottom half
includes only households in both 1993 and 1998 samples.
The contribution of each type of income to overall growth for each income group is shown in
Table 7. Here, the quintiles are defined in each year (1993 and 1998), as in the top half of the
Income diversification and poverty reduction in Vietnam
Page 14
previous table7. The table shows that growth in crop production is the most important source of
income growth for poor households, accounting for two-thirds of the income growth. Furthermore,
the contribution of crop production to income growth declines among higher-income households.
Growth in crop production accounts for 69 percent of the income growth among households in the
poorest group but just 14 percent of income growth among those in the richest group.
In contrast, the contribution of non-farm enterprise income to rural income growth is greatest
among higher income households. Among the poorest households, enterprise income actually
declined, so its “contribution” was –12 percent. In the middle three quintiles, the contribution of
enterprise income ranged from -13 to 39 percent. In the highest income category, growth in non-farm
enterprise income accounted for almost half of the overall income growth. Livestock and forestry
income make a greater contribution to the income growth of poor households than rich, while the
contribution of wage income seems to be greatest in the middle income categories.
Table 7. Contribution to overall income growth of each income source
by expenditure category in the rural Northern Uplands, 1993-1998
Income
source
Poorest
2
3
4
Richest
Average
----(percent of overall income growth)----Crops
69
55
63
23
14
45
Livestock
15
7
3
7
7
8
Fisheries
2
2
4
4
1
2
Forestry
12
6
7
2
2
6
Enterprise
-12
14
-13
34
49
16
Wages
5
14
20
15
3
11
Transfers
10
3
14
16
19
11
Other
0
0
2
0
5
1
Total
100
100
100
100
100
100
Source: Analysis of the 1993 and 1998 VLSS
Income diversification by sex of the head of household
Are the sources of income growth different between male- and female-headed households in
the rural Northern Uplands? As discussed earlier, female-headed households have per capita income
levels equal to or slightly above those of male-headed households, on average, and the growth in
income appears to be similar for both groups. Table 10 shows the contribution of each source to
overall income growth for male- and female-headed households. Almost half (46 percent) of the
7
Although not shown here, the results are very similar if the income categories are defined in terms of
the 1993 quintile, using only households interviewed in both years.
Income diversification and poverty reduction in Vietnam
Page 15
income growth of male-headed households can be attributed to growth in crop production, compared
to just 33 percent among female-headed households. The other difference between them is that
transfer income has grown more for female-headed households, contributing 40 percent of total
income growth. These transfers are mainly remittances from family members (who may include a
husband) working elsewhere. By contrast, for male-headed households, growth in transfers represents
just 7 percent of the total.
Table 8. Contribution to overall income growth of each income source
by gender of head of household in the rural Northern Uplands,
1993-1998
Income
source
Head of household
Male
Female
Average
(percent of overall income growth)
Crops
46
33
45
Livestock
8
4
8
Fisheries
1
9
2
Forestry
6
5
6
Enterprise
19
-10
16
Wages
11
12
11
Transfers
7
40
11
Other
0
8
1
Total
100
100
100
Source: Analysis of the 1993 and 1998 VLSS
Income diversification by ethnicity of the head of household
Have income growth patterns differed between Kinh/Hoa households and ethnic minority
households in the rural Northern Uplands? Earlier in this chapter, we showed that ethnic minority
households tend to be poorer than average and that the growth in their income has also been below
average. Here, we compare the composition of the income growth between 1993 and 1998. Growth
in crop income accounts for three-quarters of the income growth of ethnic minority households.
Forestry and wages are also important, each contributing 10-12 percent of the total. Enterprise
income has declined, resulting in a negative contribution (see Table 9).
In contrast, crop income barely contributed one-quarter of the total income growth for
Kinh/Hoa households. The largest contributor to income growth was enterprise income, which
accounted for over one-third (34 percent) of the total. Forestry is much less important as a source of
income growth for these households compared to ethnic minority households.
Income diversification and poverty reduction in Vietnam
Page 16
Table 9. Contribution to overall income growth of each income source
by ethnicity in the rural Northern Uplands, 1993-1998
Income
source
Ethnicity of head of household
Kinh/Hoa
Minority
Average
(percent of overall income growth)
Crops
26
74
45
Livestock
8
7
8
Fisheries
2
3
2
Forestry
2
12
6
Enterprise
34
-13
16
Wages
12
10
11
Transfers
14
7
11
Other
2
0
1
Total
100
100
100
Source: Analysis of the 1993 and 1998 VLSS
5.2
Crop diversification
The previous section compared the contribution of crop production and other economic
activities to overall rural income growth. This section focuses on the composition of the growth in
crop income.
Crop diversification in the Northern Uplands
According to the Vietnam Living Standards Surveys, the net revenue from crop production
among rural farmers in the Northern Uplands increased by about 2.0 million VND/farm/year in real
terms between 1993 and 1998.8 This section shows the composition of this growth by crop and by
source of growth: area expansion, higher prices, yield improvement, and diversification into highervalue crops. The calculations for this analysis are explained in Section 2.4.
The last column in Table 10 shows the growth in net income from different crops between the
two VLSS surveys. The growth in net income from rice was 602 thousand VND per household per
year, about 30 percent of the overall increase in net income from crops. Sugarcane and litchi/longan
each contributed another 18-20 percent of the overall increase in net income from crop production. No
other crop represents more than 6 percent of the total growth in crop income.
The other way to disaggregate the growth in crop income is by the source of the growth:
overall area increase, increased prices, higher yields, and diversification toward higher-value crops.
8
The income figures in this section differ somewhat from the ones presented in Table 3 for two
reasons. First, the sample for this analysis is smaller, being restricted to rural households in the Northern
Uplands who grow crops rather than all rural households in the Northern Uplands. Second, crop production was
defined earlier to include by-products such as straw, hay, stems, and leaves. Since these by-products are not
linked to specific crops in the questionnaire, they were excluded from this analysis.
Income diversification and poverty reduction in Vietnam
Page 17
This decomposition is shown in each row of Table 10. For example, of the VND 602 thousand
increase in the net income from rice production, about three-quarters is due to yield increases and
slightly less than one-quarter is due to area increases (this refers to sown area, so it includes cropping
intensification).
Price increases did not contribute much to the growth in the value of rice production, but it
did explain much of the growth in value of sweet potatoes, cassava, and sugar cane. In the case of
sugar cane, the higher prices are due to the government policy to achieve sugar self-sufficiency by
restricting imports, which has raised the domestic price of sugar (and indirectly sugarcane) far above
the international price.
Table 10. Composition of growth in crop income in the rural Northern Uplands, 1993-1998
Crop
Area
Price
Higher
Crop diver- Total
Expansion
Increases yields
sification
change
---------(1000 VND/household/year)-----------Rice
147
19
454
-50
602
Maize
29
42
78
-78
48
Sweet potatoes
5
114
-3
-16
66
Potatoes
1
9
-2
6
17
Cassava
16
107
-17
6
112
Other staple crops
1
-1
2
5
6
Kohlrabi, cabbage, caulif.
5
47
1
-2
55
Other leafy greens
3
18
-2
3
24
Tomatoes
2
28
-3
-6
12
Water morning glory
5
44
23
-7
80
Fresh legumes
1
3
-1
0
3
Other vegetables
3
10
-1
14
32
Soybeans
8
-6
19
19
45
Peanuts
9
-28
45
4
16
Sugar cane
11
184
-20
106
417
Tobacco
5
17
16
-17
12
Other annual crops
1
43
1
-3
3
Tea
5
0
42
-20
13
Other industrial tree crops 0
-6
-3
97
-6
Citrus
1
0
-2
11
8
Pineapple
0
0
-2
3
-1
Bananas
4
-5
3
25
29
Mango
1
-1
-8
12
-7
Apple
0
-1
41
-3
3
Plum
1
6
2
1
13
Papaya
0
2
10
-2
9
Litchi, longan & rambutan
5
24
88
35
354
Custard apple
1
1
1
8
15
Jackfruit, durian
3
10
-9
4
4
Other fruit trees
3
5
16
14
18
Total
276
685
768
169
2,002
Row percentage
14%
34%
38%
8%
100%
Source: Analysis of the 1993 and 1998 VLSS
Note: See Section 2.4 and accompanying text for explanation. Columns may not add up to total because interaction
term is not shown.
The diversification column gives the increase in the value of crop income due to reallocation
of land away from or toward that crop, holding prices, yields, and total cropped area constant.
Income diversification and poverty reduction in Vietnam
Page 18
Expansion of sugarcane, litchi/longan, and “other industrial tree crops” all represented an increase in
crop income due to crop diversification.
The bottom two rows of Table 10 shows the disaggregation of crop income by source,
summing across all crops. The largest factor in the growth of crop income in the Northern Uplands
was yield increases, which accounted for 768 VND/farm/year in additional income or 38 percent of
the total increase in crop income. In fact, yield increases in rice alone account for almost one quarter
(23 percent) of the overall increase in crop income. Price increases represented about 34 percent of
the crop income growth, while expansion in cropped area accounts for 14 percent of the total.
According to the comparison of the 1993 and 1998 VLSS studies, crop diversification increased the
average annual net revenue from crop production in the Northern Uplands by VND 169
thousand/farm. In other words, if farmers in the Northern Uplands had maintained the same total crop
area, the same yields, and the same real price, but had reallocated their land among crops following
the historical pattern between 1993 and 1998, their crop income would have increased VND 169
thousand/farm. This represents about 8 percent of the total increase in income from crop production
between the two surveys. Overall, these results indicate that, while crop diversification has
contributed to income growth in the rural Northern Uplands, it has not been as important as growth in
yields and increases in real prices.
Crop diversification in other regions
The same analysis can be carried out for the rural areas of the other regions. In the interest of
saving space, we do not present the crop-level results, but Table 11 summarizes the contribution of
each results, each row representing the In the rural areas of the other six regions, the contribution of
each factor in crop income growth. Area expansion plays a modest role in crop income growth in the
Red River Delta and the two central coast regions. In fact, the Red River Delta shows a negative
contribution, implying that the area cropped per farm household declined slightly between the two
surveys. This is not surprising given that the growth of Hanoi and the high value of land are leading
to the conversion of agricultural land to residential, industrial, and commercial uses. In contrast, area
expansion is the most important growth factor in the Southeast. Although the growth of Ho Chi Minh
Income diversification and poverty reduction in Vietnam
Page 19
City is also leading to conversion of farmland, the sown area per farm household has still increased.
Increased cropping intensity probably accounts for much of this growth in sown area.
Table 11. Sources of growth in net income from crop production by region, 1993-1998
Region
N Uplands
Red R Delta
NC Coast
SC Coast
C Highlands
Southeast
Mekong Delta
Average
Areas
expansion
14
-21
3
6
17
76
33
17
Price
Higher
increase yield
34
38
85
53
59
30
50
71
41
34
53
-12
32
44
50
39
Crop
diversification
8
13
7
4
17
8
39
17
Interaction
7
-30
1
-31
-9
-25
-48
-23
Total
100
100
100
100
100
100
100
100
Source: Analysis of the 1993 and 1998 VLSSs.
On a national level, crop diversification accounted for 17 percent of the growth in crop
income between 1993 and 1998. The contribution of crop diversification to crop income growth is
highest in the Mekong River Delta. This is not surprising given that farmers there are in the process
of reallocating land from rice to the cultivation of fruit and other higher-value commercial crops for
export and sale to Ho Chi Minh City. At the national level, yield increases represented 39 percent of
the growth, and higher real prices contributed 50 percent.
Crop diversification by income group
The growth rate in crop income does not vary in a consistent way with the level of per capita
expenditure in 1993. In other words, poor households in the rural Northern Uplands experienced as
much growth in crop production income as higher income households in that region. The
composition of this growth does, however, vary across income groups. Among the poorest farmers,
the increase in yields represents about 61 percent of the increase in crop income (see Table 12).
The contributions of area expansion, yield increases, higher prices, and crop diversification to
crop income growth in the rural Northern Uplands shows a fairly erratic pattern. Area expansion
seems to have played a more important role in crop income growth among the households that had
relatively high income. This result suggests that households with relatively high incomes in 1993
were able to use those resources to secure more land for planting crops, either through the land
allocation process, through the (formal or informal) purchase of land-use certificates, or through land
rental. The contribution of yield increases, though inconsistent, seems to indicate that this factor
Income diversification and poverty reduction in Vietnam
Page 20
plays a more important role in the crop income growth of poor rural households. This result is
plausible since yields can be increased by applying labor more intensively and through the use of
improved seed and fertilizer, which are generally scale neutral forms of agricultural technology9.
Crop diversification may play a somewhat greater role in the crop income growth of higher income
rural households, though again the pattern is not consistent. The most we can say is that the average
contribution of diversification is greater in the upper two quintiles than it is in the lower two quintiles.
Table 12. Sources of growth in net income from crop production by expenditure category
Crop
Expenditure
Areas
Price
Higher
diversicategory
expansion increase
yield
fication
Poorest
-1
29
61
10
2
8
27
54
3
3
-30
57
83
33
4
25
48
34
-1
Richest
60
34
24
24
Average
18
39
45
13
Source: Analysis of the 1993 and 1998 VLSSs (panel households only).
Note: Expenditure categories are defined according to the level in 1993.
Interaction
2
9
-44
-6
-32
-15
Total
100
100
100
100
100
100
These patterns are also seen across quintiles in other regions. Combining all the rural farm
households together and classifying them by expenditure quintile, we see that crop diversification and
area expansion contribute more to crop income growth among higher income households than among
poor households. Furthermore, as in the Northern Uplands, poor households rely more on yield
increases to boost the value of their crop income. On a national level, crop diversification accounts
for 27 percent of the growth in crop income in our sample of rural farm households that were in both
the 1993 and 1998 VLSS surveys10.
6
Summary and discussion
The comparison of the two surveys indicates substantial growth in income and expenditure
over 1993-98. The gains for rural households have been widespread across regions and types of
households. In the rural Northern Uplands, per capita expenditure grew 31 percent between 1993 and
1998, roughly equal to the national average for rural households. Although the gap between poor and
9
Because seed and fertilizer are divisible, this type of agricultural technology is considered more
scale- neutral than mechanical technology.
10
When we consider the full samples of rural households in each survey, the contribution of crop
diversification to crop income growth was 22 percent, as discussed above.
Income diversification and poverty reduction in Vietnam
Page 21
rich households widened between 1993 and 1998, paradoxically those who were poor in 1993 gained
as much on average as those who were “rich” in 1993. Thus, there is little evidence that the rural poor
have, in general, gotten poorer or been left behind in the rise in standards of living over the 1990s.
At the same time, crop production continues to be the most important source of income for
rural households, accounting for 46 percent of the net income in the Northern Uplands. Poor rural
households depend even more on crop income than other rural households. Staple food crops,
particularly rice, continue to play a dominant role in crop production. Rice alone accounts for 46
percent of the net value of crop production.
Non-crop income among rural households in the Northern Uplands represented 54 percent of
the total, so it is important but its share did not grow over the period 1993-98. This implies that
growth in non-crop income accounted for slightly over half of total growth, but the growth rates of
crop and non-crop income were almost identical. Similarly, non-agricultural income was stable at
about 37 percent of the total. National accounts statistics indicate that, over 1995-2000, the nonagricultural sector grew from 53 to 58 percent of regional gross domestic product. If the VLSS
estimates are accurate, then the expansion in the relative expansion in the non-agricultural sector may
be limited to urban areas.
Over 1993-1998, the growth in crop income accounted for 45 percent of the growth in overall
income for the average rural household in the Northern Uplands, but crop income contributes an even
higher percentage among the poorest rural households (69 percent) and among ethnic minority
households (74 percent).
Decomposing crop income growth, 38 percent is attributable to higher yields, 34 percent to
higher real prices, 14 percent to area expansion, and 8 percent to crop diversification. The crop
diversification in the northern uplands is mainly in the form of farmers reducing the area planted with
rice and increasing the area planted to sugarcane and fruit. Nationally, crop diversification accounts
for 17 percent of the growth in crop income. The contribution of crop diversification is highest in the
Mekong Delta, where farmers have converted rice land to expand production of fruit and aquaculture.
The sources of crop income growth vary across income groups. Poor households increased
their crop income largely by achieving higher yields, particularly for rice, while richer households
Income diversification and poverty reduction in Vietnam
Page 22
increased their incomes by expanding the area cultivated. The contribution of diversification shows
an erratic pattern across income categories, but appears to be less important for poor rural households
than others.
Four implications can be drawn from these results. First, national economic growth has been
translated into rapid improvements in standards of living in the poorest area of rural Vietnam.
Although the income gap is widening, there is no general tendency for the poor to get poorer, be left
behind, or even experience lower-than-average growth rates. Second, this growth has been based
largely on growth in crop production, in particular from price and yield increases. This suggests that
crop production is not necessarily a “dead-end” for small farmers. Third, assuming that resources
should be allocated toward factors in proportion to their contribution to growth, priority should be
given to agricultural research and extension in the northern uplands. This does not imply that
programs to foster diversification should be eliminated, but they should not be allowed to divert
resources from technology development. Fourth, the importance of crop diversification varies widely
by region, being much more important among farmers in the Mekong Delta. Finally, since income
from wages and non-farm enterprises is positively correlated with household income, untargeted
efforts to assist existing non-farm enterprises will generally favor high-income rural households.
Income diversification and poverty reduction in Vietnam
Page 23
References
Barrett, C., M. Bezuneh, and A. Aboud. 2001. “Income diversification, poverty traps, and policy
shocks in Cote d’Ivoire and Kenya.” Food Policy 26: 367-384.
Barrett, C., M. Bezuneh, D. Clay, and T. Reardon. 2000. “Heterogeneous constraints, incentives, and
income diversification strategies in rural Africa.” Broadening Access and Strengthening
Input Market Systems. University of Wisconsin, Madison, Wisconsin.
Delgado, C. and A. Siamwalla. 1997. “Rural economy and farm income diversification in developing
countries.” Discussion Paper No. 20. Markets and Structural Studies Division. International
Food Policy Research Institute. Washington, D.C.
Delgado, C., J. Hopkins, and V. Kelly with P. Hazell, A. McKenna, P. Gruhn, B. Hojjati, J. Sil, and C.
Courbois. 1999. Agricultural Growth Linkages in Sub-Saharan Africa. Research Report No.
107. International Food Policy Research Institute. Washington, D.C.
Escobal, J. 2001. “The determinants of nonfarm income diversification in rural Peru.” World
Development 29 (3): 497-508.
General Statistical Office. 2000a. Vietnam Living Standards Survey 1997-1998. Statistical Publishing
House, Hanoi.
Hazell, P. and A. Roell. 1983. Rural growth linkages: Household expenditure patterns in Malaysia and
Nigeria. Research Report 41. International Food Policy Research Institute. Washington, D.C.
Henin, B. 2002. “Agrarian change in Vietnam’s northern upland region.” Journal of Contemporary
Asia. 32 (1): 3-28.
Joint Working Group. 2000. Vietnam Development Report 2000: Attacking Poverty. Joint report of
the Government-Donor-NGO Working Group. Hanoi, Vietnam.
Joshi, P.K, A. Gulati, P.S. Birthal, and L. Twari. 2002. “Agricultural diversification in South Asia:
Patterns, determinants, and policy implications.” Forthcoming Discussion Paper. Markets and
Structural Studies Division. International Food Policy Research Institute. Washington, D.C.
Lanjouw, J. and P. Lanjouw. 2001. “The rural non-farm sector: Issues and evidence from developing
countries.” Agricultural Economics 26: 1-13.
Minot, N. and B. Baulch. 2002. “The Spatial Distribution of Poverty in Vietnam and the Potential for
Targeting.” Discussion Paper No. 43. Markets and Structural Studies Division. International
Food Policy Research Institute. Washington, D.C.
Pederson, G. and M. Annou. 1999. “Rural household income diversification in Viet Nam.” Quarterly
Journal of International Agriculture. 2.
Reardon, T. 1997. “Using evidence of household income diversification to inform the study of rural
nonfarm labor market in Africa.” World Development 25 (5): 735-747.
State Planning Commission/General Statistics Office. 1994. Vietnam Living Standards Survey 199293. State Planning Committee and General Statistics Office. Hanoi.
Income diversification and poverty reduction in Vietnam
Page 24